In cooperation with a Dutch telecom provider, Statistics Netherlands has been developing a model for the tracking of mobile devices. Using multiple observations of the device of an employee, a probability distribution of the traveled path can be computed. This article serves as a proof of concept for the proposed methods. The available data contains records of connections between antennas and the device. A separate data source contains properties of all antennas in the network. A signalling strength model translates the properties of each antenna into a probability distribution of the position of a device in a 100m by 100m grid cell. A Markov Chain Monte Carlo model uses these probability distributions as input in the form of observations. The proposed approach does not only result in an estimated travel path of the device, it also provides the estimation of transportation mode for each time step.